Skip to main content

Limiting factors in high performance computing

  • Conference paper
  • First Online:
Parallel Scientific Computing (PARA 1994)

Part of the book series: Lecture Notes in Computer Science ((LNCS,volume 879))

Included in the following conference series:

Abstract

By examining the details of typical HPC computations, we have shown that data output rates represent the major obstacle to scaling HPC computations to the Tflops range. If past experience is to be a guide, we can safely conclude that manufacturers will concentrate on peak performance and internal I/O rates while largely ignoring the more significant external I/O issues. Unless pressure is put on the vendors, there is a real danger that Tflops machines will not deliver on expectations for this reason. High speed communication protocols and switching systems such as a ATM will likely play an increasingly important role in HPC in coming years.

Invited talk at the PARA94 Conference, Copenhagen, Denmark, June 21–24, 1994.

To appear in Proceedings of the PARA94 Conference, ed. J. Wazniewski, Lectures in Computer Science, Springer-Verlag, 1994.

Research supported in part by NSF Grand Challenges Application Group grant ASC-9217394 and by NASA HPCC Group grant NAG5-2218.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Institutional subscriptions

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. F. Cattaneo, N. H. Brummell, J Toomre, A. Malagoli, and N. E. Hurlburt, Astrophys. Jou., 370, 282,1992.

    Google Scholar 

  2. C. F. Baillie and O. McBryan, “Computer Science Aspects of Grand Challenge Computing”, in Proceedings of High Performance Computing '94, ed. A. Tentner, pp. 247–251, The Society for Computer Simulation, San Diego, 1994.

    Google Scholar 

  3. C. Farhat and O. McBryan, “Heterogeneous Computing for Aeroelastic Computations”, Proceedings of the Seventh SIAM Conference on Parallel Processing for Scientific Computing, SIAM, to appear.

    Google Scholar 

  4. J. Blondeau, R. G. Foldvick and O. McBryan, “BATMAN: A Wide Area ATM Network”, Proceedings ICC '95, IEEE to appear.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Editor information

Jack Dongarra Jerzy Waśniewski

Rights and permissions

Reprints and permissions

Copyright information

© 1994 Springer-Verlag Berlin Heidelberg

About this paper

Cite this paper

McBryan, O.A. (1994). Limiting factors in high performance computing. In: Dongarra, J., Waśniewski, J. (eds) Parallel Scientific Computing. PARA 1994. Lecture Notes in Computer Science, vol 879. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0030165

Download citation

  • DOI: https://doi.org/10.1007/BFb0030165

  • Published:

  • Publisher Name: Springer, Berlin, Heidelberg

  • Print ISBN: 978-3-540-58712-5

  • Online ISBN: 978-3-540-49050-0

  • eBook Packages: Springer Book Archive

Publish with us

Policies and ethics